Combining colour and grey‐level co‐occurrence matrix features: a comparative study

Abstract

Grey‐level co‐occurrence matrix (GLCM) is a widely used texture feature descriptor that is extracted from grey‐level images. A considerable amount of work in the literature has been done trying to combine colour with GLCM features. The aim of this study is to examine the effect of integrating colour with GLCM. To this end, 14 descriptors that have been proposed in literature based on different colour combining approaches, four Convolutional Neural Networks‐based, two aggregation‐based, and five other handcrafted descriptors have been engaged in this comparison. Five widely known data sets have been used to experimentally evaluate each descriptor. The evaluation has been conducted for image classification and image retrieval. Results indicated that there is no co‐occurrence‐matrix‐based descriptor that overcomes all the others in all data sets. Nevertheless, some descriptors have shown a remarkably higher performance than the others.